Zero Shot Molecular Generation via Similarity Kernels

Abstract

Generative modelling aims to accelerate the discovery of novel chemicals by directly proposing structures with desirable properties. Recently, score-based, or diffusion, generative models have significantly outperformed previous approaches. Key to their success is the close relationship between the score and physical force, allowing the use of powerful equivariant neural networks. However, the behaviour of the learnt score is not yet well understood. Here, we analyse the score by training an energy-based diffusion model for molecular generation. We find that during the generation the score resembles a restorative potential initially and a quantum-mechanical force at the end. In between the two endpoints, it exhibits special properties that enable the building of large molecules. Using insights from the trained model, we present Similarity-based Molecular Generation (SiMGen), a new method for zero shot molecular generation. SiMGen combines a time-dependent similarity kernel with descriptors from a pretrained machine learning force field to generate molecules without any further training. Our approach allows full control over the molecular shape through point cloud priors and supports conditional generation. We also release an interactive web tool that allows users to generate structures with SiMGen online (https://zndraw.icp.uni-stuttgart.de).

References in corpus (36)

• Denoising Diffusion Probabilistic Models
• CP2K: An Electronic Structure and Molecular Dynamics Software Package -- Quickstep: Efficient and Accurate Electronic Structure Calculations
• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• On representing chemical environments
• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
• Deep Unsupervised Learning using Nonequilibrium Thermodynamics
• Score-Based Generative Modeling through Stochastic Differential Equations
• Ab initio Random Structure Searching
• Generative Modeling by Estimating Gradients of the Data Distribution
• High-pressure phases of silane
• DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking
• Elucidating the Design Space of Diffusion-Based Generative Models
• MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields
• MoFlow: An Invertible Flow Model for Generating Molecular Graphs
• GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation
• Equivariant Diffusion for Molecule Generation in 3D
• Crystal Diffusion Variational Autoencoder for Periodic Material Generation
• Structure-based Drug Design with Equivariant Diffusion Models
• MatterGen: a generative model for inorganic materials design
• MACE-OFF: Transferable Short Range Machine Learning Force Fields for Organic Molecules
• H5MD: a structured, efficient, and portable file format for molecular data
• DiffBP: Generative Diffusion of 3D Molecules for Target Protein Binding
• Pre-training via Denoising for Molecular Property Prediction
• GraphEBM: Molecular Graph Generation with Energy-Based Models
• Diffusion-based Molecule Generation with Informative Prior Bridges
• An Overview of Diffusion Models: Applications, Guided Generation, Statistical Rates and Optimization
• MUDiff: Unified Diffusion for Complete Molecule Generation
• SPICE, A Dataset of Drug-like Molecules and Peptides for Training Machine Learning Potentials
• Two for One: Diffusion Models and Force Fields for Coarse-Grained Molecular Dynamics
• Navigating the Design Space of Equivariant Diffusion-Based Generative Models for De Novo 3D Molecule Generation
• Learning Interpolations between Boltzmann Densities
• Compositional Visual Generation with Composable Diffusion Models
• Inversion of the chemical environment representations
• GenMol: A Drug Discovery Generalist with Discrete Diffusion
• ZnTrack -- Data as Code